Shaokai Zhang1, Leni Kang1, Bin Liu1, Jianfeng Cui1, Feng Chen1, Xinfu Liu1, Hong Wang1, Wen Chen2. 1. Department of Cancer Epidemiology, Cancer Institute/Hospital, Chinese Academy of Medical Sciences and Peking Union Medical College, Beijing 100021, China. 2. Department of Cancer Epidemiology, Cancer Institute/Hospital, Chinese Academy of Medical Sciences and Peking Union Medical College, Beijing 100021, China. Email: chenwen@cicams.ac.cn.
Abstract
OBJECTIVE: To evaluate the diagnostic performance of different specimens for detecting CIN2(+), and to find the solution of the problem that why the performance of self-collected specimen is worse than cervical specimen collected by physician. METHODS: The cervix, lower 1/3 vagina, upper 1/3 vagina and self-collected specimens from each of the 806 women who took part in this multi-center screening program from May 2006 to April 2007 were tested by hybrid capture 2 (HC2) technique. The diagnostic performance of HC2 on the four specimens for detecting CIN2(+) lesions was calculated. Linear array was performed on the four specimens from 489 out of the 806 women and the diagnostic performance of linear array on the four specimens for detecting CIN2(+) lesions was also calculated. Z test was used to compare the area under ROC and McNemar or χ(2) test was used to compare the sensitivity and specificity of different specimens. RESULTS: The area under ROC of the cervix, 1/3 upper vagina, 1/3 lower vagina and self-collected samples testing by HC2 for detecting CIN2(+) lesions were 0.902, 0.793, 0.769 and 0.773, respectively (P < 0.001). Using 1 RUL/CO as the cut-point of HC2, the sensitivity of the cervix, upper vagina, lower vagina and self-collected samples were 98.0%, 91.8%, 83.7% and 81.6%. Compared with the cervical specimen, the sensitivity of self-collected specimen for detecting CIN2(+) lesions was significantly lower (P = 0.008). Lowering the cutoff value for HC2 test could improve the sensitivity of self-collected specimen, but it significantly compromised the specificity. The sensitivity of self-collected specimen tested by linear array for detecting CIN2(+) lesions was 95.7% and it was not significantly different compared with the sensitivity of cervical specimen (97.9%) tested by HC2. CONCLUSIONS: The performance of self-collected specimen tested by HC2 for detecting CIN2(+) lesions is lower than that of physician-collected cervical specimen, and lowering the cutoff value can't improve its diagnostic performance. Using linear array as the HPV DNA test can significantly improve the screening diagnostic performance of self-collected specimens.
OBJECTIVE: To evaluate the diagnostic performance of different specimens for detecting CIN2(+), and to find the solution of the problem that why the performance of self-collected specimen is worse than cervical specimen collected by physician. METHODS: The cervix, lower 1/3 vagina, upper 1/3 vagina and self-collected specimens from each of the 806 women who took part in this multi-center screening program from May 2006 to April 2007 were tested by hybrid capture 2 (HC2) technique. The diagnostic performance of HC2 on the four specimens for detecting CIN2(+) lesions was calculated. Linear array was performed on the four specimens from 489 out of the 806 women and the diagnostic performance of linear array on the four specimens for detecting CIN2(+) lesions was also calculated. Z test was used to compare the area under ROC and McNemar or χ(2) test was used to compare the sensitivity and specificity of different specimens. RESULTS: The area under ROC of the cervix, 1/3 upper vagina, 1/3 lower vagina and self-collected samples testing by HC2 for detecting CIN2(+) lesions were 0.902, 0.793, 0.769 and 0.773, respectively (P < 0.001). Using 1 RUL/CO as the cut-point of HC2, the sensitivity of the cervix, upper vagina, lower vagina and self-collected samples were 98.0%, 91.8%, 83.7% and 81.6%. Compared with the cervical specimen, the sensitivity of self-collected specimen for detecting CIN2(+) lesions was significantly lower (P = 0.008). Lowering the cutoff value for HC2 test could improve the sensitivity of self-collected specimen, but it significantly compromised the specificity. The sensitivity of self-collected specimen tested by linear array for detecting CIN2(+) lesions was 95.7% and it was not significantly different compared with the sensitivity of cervical specimen (97.9%) tested by HC2. CONCLUSIONS: The performance of self-collected specimen tested by HC2 for detecting CIN2(+) lesions is lower than that of physician-collected cervical specimen, and lowering the cutoff value can't improve its diagnostic performance. Using linear array as the HPV DNA test can significantly improve the screening diagnostic performance of self-collected specimens.